. This application seeks continued support for the Study of Dementia in Swedish Twins to expand the sample by means of total ascertainment of all cases of Alzheimer's disease and other late-life dementias in the entire Swedish Twin Registry. The estimate is to complete data for over 200 pairs in which one or both is diagnosed with Alzheimer's disease, and 350 pairs in which one or both has any dementia diagnosis. Case identification will use telephone screening with all individuals in the sample, followed by informant interviews for those with evidence impairment, resulting in cognitive screening data from over 6000 complete pairs. Record linkage to health care utilization will be carried out as a parallel case identification strategy. Diagnostic assessment for those with a positive screening outcome will include medical evaluation, neuropsychological measures, neuroimaging, and informant interviews. Partners will receive the identical protocol, as will an independent sample of individuals who screened negative. Risk and protective factors will be obtained from the informant interview, medical records, and twin registry database, with occupational history linked to measured exposures. One longitudinal follow-up is included, and permission for autopsy will be requested. Data analyses using quantitative genetic approaches will address six questions: 1) What is the relative importance of genetic and environmental effects of Alzheimer's disease? 2) What are the best indices of the underlying genetic liability and of quantitative variability? 3) What non-genetic risk factors can be identified? 4) What influences age of onset? 5) Are there interactions between genetic and environmental risk factors? and 6) What is the covariation of liability to different types of disorder? Analytic techniques will include expansion of liability threshold models to mixed models, sex limitation models, use of measured genotype, bivariate models, and case-control methods using matched twin designs. Different diagnostic criteria will be compared. Quantitative indices will be derived from neuropsychological scores and from neuroimaging data subjected to quantitative mapping.